--- title: "Your Workflow, Turned Into an Agent Your Team Owns" slug: "process-workflow-to-agent-video" date: "2026-11-08" tag: "Application" summary: "A 55-second clip of the ibl.ai process page, ibl.ai/process: a department describes one workflow in its own words, a person replies within one business day, a 30-minute call sets the agent's limits, engineers build on a safe copy of the data, the team writes the test set, and the agent runs in the team's own cloud on code it owns." author: "ibl.ai Engineering" linkedin: | A department can hand one workflow to an agent without a data project or a line of code from its own staff. We put the whole process on one page, ibl.ai/process, and this 55-second clip walks through it. → You describe the workflow in your own words, including the number that will show it worked → A person replies within one business day, and a 30-minute call settles what the agent may and may not do → Our engineers build it on a safe copy of your data, through connectors scoped to each person's role → Your team writes the test set, and a version that scores below the bar does not go live → Then you run it in your own cloud, and the skills are plain-English files your team edits without waiting on us → The connectors and the test set carry over, so the second workflow starts further along than the first With ibl.ai you own all the code and the data: self-hosted inside your own perimeter, model-agnostic across any LLM, with no per-seat pricing, deployable anywhere from your own cloud to a fully air-gapped network. #iblai #AgenticAI #EnterpriseAI #AIAgents #SovereignAI --- **ibl.ai turns a department's everyday workflow into an AI agent from a plain-English description, and the new [ibl.ai/process](/process) page shows the whole path in seven stages: you describe one workflow, a person replies within one business day, a 30-minute call sets the agent's limits, our engineers build it on a safe copy of your data, your team writes the test set, and the agent runs in your own cloud, where with ibl.ai you own all the code and the data.** The 55-second clip above walks the page and the platform in order: the four sections of the form, the seven stages, then the Integrations, Evals and Skills panels of an agent on the Lakeview University demo organization, recorded on November 8, 2026. The page is written for department leads and administrators, not engineers. Filling in the form takes about 10 minutes, needs no technical knowledge, and covers one workflow at a time. ## What is on the ibl.ai/process page? [ibl.ai/process](/process) puts everything a department needs before it hands a workflow to an agent on one page: the **7 stages** from the form to the next workflow, the **4 sections** of the form with an example answer beside each, the **13 questions** a team asks before an agent joins it, the **5 routes** to a system's data, and **6 good first agents**, one per department. Every stage names who takes part. The person who does the work today describes it, tries the first version and approves what goes live. The form itself is embedded at the foot of the page, and the second ask on [AI Transformation](/service/ai-transformation), "Turn your workflow into an agent", now lands there, as does the band on every Solutions page. ## What happens after a department submits a workflow? A person reads every submission and replies within **one business day** with two or three follow-up questions or a short call invitation. A **30-minute call** with the team and whoever owns the systems settles what the agent should and should not do, writes down the number that will show it worked, and names the person on the department's side who will run it. Our engineers then build the first version on a safe copy of the data, reading each system in place through [Agentic Ontology](/product/agentic-ontology) connectors scoped to each person's role. Nothing is copied into a warehouse first. ## How does the team know the agent is right before it goes live? The team writes a test set: questions with the answers it agrees are correct. Every version of the agent is scored against it, and a version below the bar does not go live. The mechanism is the one shipped as [agent evals](/updates/agent-evals-benchmarks-llm-judges-csv-export), with the benchmark visible in the clip. Every step the team marks for approval stays with a person. The agent acts on its own only where the department says it may. ## What does the department keep when the agent runs? The agent runs in the department's own cloud, on-premise or fully air-gapped, on any model, and with ibl.ai you own all the code and the data. Skills are plain-English files the team edits without waiting on us, following the [eight-step agentic development process](/architecture#agentic-development). The connectors, the test set and the skills carry over, so the second workflow starts further along than the first, and the team can lead it. ## Where does the clip go from here? The clip sits at the top of [ibl.ai/process](/process) under the two asks, plays muted inline, and opens narrated with English captions on a click. The master is staged for a YouTube upload; this page will link the public copy once it exists. To start, describe one workflow at [ibl.ai/process](/process), or [book a 30-minute call](https://cal.com/iblai/30min) and describe it out loud; we take the notes.